Record language usage as UMA histograms.

Record accept languages and application language as UMA histograms on
browser start-up.

See also http://crosbug.com/17419

BUG=none
TEST=run the unit tests


Review URL: http://codereview.chromium.org/7348004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@93360 0039d316-1c4b-4281-b951-d872f2087c98

6 files changed, 215 insertions, 0 deletions

diff --git a/chrome/browser/browser_main.cc b/chrome/browser/browser_main.cc
index 3a17cb0..d33e4a3 100644
--- a/chrome/browser/browser_main.cc
+++ b/chrome/browser/browser_main.cc
@@ -45,6 +45,7 @@
 #include "chrome/browser/first_run/first_run_browser_process.h"
 #include "chrome/browser/first_run/upgrade_util.h"
 #include "chrome/browser/jankometer.h"
+#include "chrome/browser/language_usage_metrics.h"
 #include "chrome/browser/metrics/field_trial_synchronizer.h"
 #include "chrome/browser/metrics/histogram_synchronizer.h"
 #include "chrome/browser/metrics/metrics_log.h"
@@ -1883,6 +1884,10 @@ int BrowserMain(const MainFunctionParams& parameters) {
   HandleTestParameters(parsed_command_line);
   RecordBreakpadStatusUMA(metrics);
   about_flags::RecordUMAStatistics(local_state);
+  LanguageUsageMetrics::RecordAcceptLanguages(
+      profile->GetPrefs()->GetString(prefs::kAcceptLanguages));
+  LanguageUsageMetrics::RecordApplicationLanguage(
+      g_browser_process->GetApplicationLocale());
 
 #if defined(OS_CHROMEOS)
   metrics->StartExternalMetrics();
diff --git a/chrome/browser/language_usage_metrics.cc b/chrome/browser/language_usage_metrics.cc
new file mode 100755
index 0000000..b53c25f
--- /dev/null
+++ b/chrome/browser/language_usage_metrics.cc
@@ -0,0 +1,67 @@
+// Copyright (c) 2011 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/language_usage_metrics.h"
+
+#include <algorithm>
+
+#include "base/metrics/histogram.h"
+#include "base/string_tokenizer.h"
+#include "base/string_util.h"
+
+namespace {
+void RecordAcceptLanguage(Language language) {
+  UMA_HISTOGRAM_ENUMERATION("LanguageUsageMetrics.AcceptLanguage",
+                            language, NUM_LANGUAGES);
+}
+}  // namespace
+
+// static
+void LanguageUsageMetrics::RecordAcceptLanguages(
+    const std::string& accept_languages) {
+  // Rethink about the histogram memory costs when the number of languages
+  // becomes too large.
+  DCHECK_LE(NUM_LANGUAGES, 300);
+
+  std::set<Language> languages;
+  ParseAcceptLanguages(accept_languages, &languages);
+  std::for_each(languages.begin(), languages.end(), RecordAcceptLanguage);
+}
+
+// static
+void LanguageUsageMetrics::RecordApplicationLanguage(
+    const std::string& application_locale) {
+  const Language language = ToLanguage(application_locale);
+  if (language != UNKNOWN_LANGUAGE) {
+    UMA_HISTOGRAM_ENUMERATION("LanguageUsageMetrics.ApplicationLanguage",
+                              language, NUM_LANGUAGES);
+  }
+}
+
+// static
+void LanguageUsageMetrics::ParseAcceptLanguages(
+    const std::string& accept_languages, std::set<Language>* languages) {
+  languages->clear();
+  StringTokenizer locales(accept_languages, ",");
+  while (locales.GetNext()) {
+    const Language language = ToLanguage(locales.token());
+    if (language != UNKNOWN_LANGUAGE) {
+      languages->insert(language);
+    }
+  }
+}
+
+// static
+Language LanguageUsageMetrics::ToLanguage(const std::string& locale) {
+  StringTokenizer parts(locale, "-_");
+  if (!parts.GetNext()) {
+    return UNKNOWN_LANGUAGE;
+  }
+  const std::string language_part = parts.token();
+  Language language;
+  if (!LanguageFromCode(language_part.c_str(), &language)) {
+    return UNKNOWN_LANGUAGE;
+  }
+  return language;
+}
diff --git a/chrome/browser/language_usage_metrics.h b/chrome/browser/language_usage_metrics.h
new file mode 100755
index 0000000..e5cc708
--- /dev/null
+++ b/chrome/browser/language_usage_metrics.h
@@ -0,0 +1,48 @@
+// Copyright (c) 2011 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef CHROME_BROWSER_LANGUAGE_USAGE_METRICS_H_
+#define CHROME_BROWSER_LANGUAGE_USAGE_METRICS_H_
+#pragma once
+
+#include <set>
+#include <string>
+
+#include "base/gtest_prod_util.h"
+#include "third_party/cld/languages/public/languages.h"
+
+// Methods to record language usage as UMA histograms.
+// Language codes are defined in third_party/cld/languages/proto/languages.pb.h
+class LanguageUsageMetrics {
+ public:
+  // Records accept languages as a UMA histogram. |accept_languages| is a
+  // case-insensitive comma-separated list of languages/locales of either xx,
+  // xx-YY, or xx_YY format where xx is iso-639 language code and YY is iso-3166
+  // country code. Country code is ignored. That is, xx and XX-YY are considered
+  // identical and recorded once.
+  static void RecordAcceptLanguages(const std::string& accept_languages);
+
+  // Records the application language as a UMA histogram. |application_locale|
+  // is a case-insensitive locale string of either xx, xx-YY, or xx_YY format.
+  // Only the language part (xx in the example) is considered.
+  static void RecordApplicationLanguage(const std::string& application_locale);
+
+ private:
+  // This class must not be instantiated.
+  LanguageUsageMetrics();
+
+  // Parses |accept_languages| and returns a set of language codes in
+  // |languages|.
+  static void ParseAcceptLanguages(const std::string& accept_languages,
+                                   std::set<Language>* languages);
+
+  // Parses |locale| and returns the language code. Returns UNKNOWN_LANGUAGE in
+  // case of errors.
+  static Language ToLanguage(const std::string& locale);
+
+  FRIEND_TEST_ALL_PREFIXES(LanguageUsageMetricsTest, ParseAcceptLanguages);
+  FRIEND_TEST_ALL_PREFIXES(LanguageUsageMetricsTest, ToLanguage);
+};
+
+#endif  // CHROME_BROWSER_LANGUAGE_USAGE_METRICS_H_
diff --git a/chrome/browser/language_usage_metrics_unittest.cc b/chrome/browser/language_usage_metrics_unittest.cc
new file mode 100644
index 0000000..db6e987
--- /dev/null
+++ b/chrome/browser/language_usage_metrics_unittest.cc
@@ -0,0 +1,91 @@
+// Copyright (c) 2011 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/language_usage_metrics.h"
+
+#include "testing/gtest/include/gtest/gtest.h"
+
+TEST(LanguageUsageMetricsTest, ParseAcceptLanguages) {
+  std::set<Language> language_set;
+  std::set<Language>::const_iterator it;
+
+  // Basic single language case.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // Empty language.
+  LanguageUsageMetrics::ParseAcceptLanguages("", &language_set);
+  EXPECT_EQ(0U, language_set.size());
+
+  // Country code is ignored.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja-JP", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // Case is ignored.
+  LanguageUsageMetrics::ParseAcceptLanguages("Ja-jP", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // Underscore as the separator.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja_JP", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // The result contains a same language code only once.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja-JP,ja", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // Basic two languages case.
+  LanguageUsageMetrics::ParseAcceptLanguages("en,ja", &language_set);
+  EXPECT_EQ(2U, language_set.size());
+  it = language_set.begin();
+  EXPECT_EQ(ENGLISH, *it);
+  EXPECT_EQ(JAPANESE, *++it);
+
+  // Multiple languages.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja-JP,en,es,ja,en-US",
+                                             &language_set);
+  EXPECT_EQ(3U, language_set.size());
+  it = language_set.begin();
+  EXPECT_EQ(ENGLISH, *it);
+  EXPECT_EQ(JAPANESE, *++it);
+  EXPECT_EQ(SPANISH, *++it);
+
+  // Two empty languages.
+  LanguageUsageMetrics::ParseAcceptLanguages(",", &language_set);
+  EXPECT_EQ(0U, language_set.size());
+
+  // Trailing comma.
+  LanguageUsageMetrics::ParseAcceptLanguages("ja,", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(JAPANESE, *language_set.begin());
+
+  // Leading comma.
+  LanguageUsageMetrics::ParseAcceptLanguages(",es", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(SPANISH, *language_set.begin());
+
+  // Combination of invalid and valid.
+  LanguageUsageMetrics::ParseAcceptLanguages("zz,en", &language_set);
+  EXPECT_EQ(1U, language_set.size());
+  EXPECT_EQ(ENGLISH, *language_set.begin());
+}
+
+TEST(LanguageUsageMetricsTest, ToLanguage) {
+  // Basic case.
+  EXPECT_EQ(JAPANESE, LanguageUsageMetrics::ToLanguage("ja"));
+
+  // Case is ignored.
+  EXPECT_EQ(SPANISH, LanguageUsageMetrics::ToLanguage("Es"));
+
+  // Coutry code is ignored.
+  EXPECT_EQ(JAPANESE, LanguageUsageMetrics::ToLanguage("ja-JP"));
+
+  // Invalid locales are considered as unknown language.
+  EXPECT_EQ(UNKNOWN_LANGUAGE, LanguageUsageMetrics::ToLanguage(""));
+  EXPECT_EQ(UNKNOWN_LANGUAGE, LanguageUsageMetrics::ToLanguage("xx"));
+}
diff --git a/chrome/chrome_browser.gypi b/chrome/chrome_browser.gypi
index 9a9c605..6227d54 100644
--- a/chrome/chrome_browser.gypi
+++ b/chrome/chrome_browser.gypi
@@ -37,6 +37,7 @@
         '../printing/printing.gyp:printing',
         '../skia/skia.gyp:skia',
         '../third_party/bzip2/bzip2.gyp:bzip2',
+        '../third_party/cld/cld.gyp:cld',
         '../third_party/expat/expat.gyp:expat',
         '../third_party/hunspell/hunspell.gyp:hunspell',
         '../third_party/icu/icu.gyp:icui18n',
@@ -1348,6 +1349,8 @@
         'browser/keychain_mac.h',
         'browser/language_combobox_model.cc',
         'browser/language_combobox_model.h',
+        'browser/language_usage_metrics.cc',
+        'browser/language_usage_metrics.h',
         'browser/language_order_table_model.cc',
         'browser/language_order_table_model.h',
         'browser/mac/authorization_util.h',
diff --git a/chrome/chrome_tests.gypi b/chrome/chrome_tests.gypi
index 114a2f7..b9e7539 100644
--- a/chrome/chrome_tests.gypi
+++ b/chrome/chrome_tests.gypi
@@ -1467,6 +1467,7 @@
         'browser/instant/instant_loader_manager_unittest.cc',
         'browser/instant/promo_counter_unittest.cc',
         'browser/internal_auth_unittest.cc',
+        'browser/language_usage_metrics_unittest.cc',
         'browser/mac/keystone_glue_unittest.mm',
         'browser/media/media_internals_unittest.cc',
         'browser/metrics/display_utils_unittest.cc',